The current tax treatment of pensions and health insurance in the United States is a hybrid that lacks consistency under either an accrual income tax system or a consumption tax system. Under an accrual income tax, employer contributions to pension plans represent an addition to wealth that would be taxed at the time they are made. The interest earned on pension contributions also represents an addition to wealth that would be taxed annually. When a worker retires, all applicable taxes would already have been paid on the benefit, and the flow of retirement income received by the worker would not be taxed. Similarly, employer-provided health insurance arguably represents a current benefit that, under the income tax, should be taxed annually as current income. Under a consumption tax, things are different for pensions. The idea of the consumption tax is to tax what an individual takes out of the system. Since pension contributions represent saving, they are not taxed when they are made. Neither is the interest earned on pension contributions taxed under a consumption tax, since it is reinvested and accumulated. Only when the worker retires and starts to draw retirement income are pension contributions taxed. And only the portion of the retirement income that is consumed is taxed if only half is consumed, taxes are paid only on that half. Although pensions fare better under a consumption tax than under an income tax, it is unclear whether health insurance would, too. If health insurance expenditures are considered current consumption (as most economists believe they should be), the same tax that applied to any other consumption would apply to employer contributions to health insurance. On the other hand, one could argue (as in footnote 3) that health insurance is a merit good and medical expenditures are unfortunate, so that both pensions and health insurance should be excluded from the definition of consumption. The existing tax treatment of employee benefits in the U.S. is a hybrid because we nominally have an income tax under which employer contributions to both pensions and health insurance could be taxed as income. But both receive favorable tax treatment pensions are tax-favored in that current pension contributions and interest on previous contributions go untaxed, and health insurance contributions are tax-free. The tax treatment of pensions is consistent with a consumption tax, not an income tax, and the prevailing view among economists is that the tax treatment of health insurance is consistent with neither. Current attempts to move toward a consumption tax have been welcomed by most economists both because most subscribe to the basic claims that are made for the consumption tax increased saving, improved economic growth, and greater efficiency and because the consumption tax promises to bring greater coherence to a system that, despite improvements during the last 15 years, still has some basic inconsistencies. Major concerns with the consumption tax have been raised by many employers, though, who are comfortable with the existing tax treatment of employee benefits and less obsessed than economists with the notion of allocative efficiency or with making the tax system conform to a consistent theory of taxation. Employers especially employers of skilled labor have at least two reasons for wanting to provide employee benefits and accordingly find the favorable tax treatment of benefits attractive (Rosen forthcoming). First, provision of employee benefits may have externalities that enhance the productivity of workers. For example, employers may want to ensure that their workers have good access to health care so that they are more likely to stay healthy. And they may want to provide pension benefits to workers to relieve workers of the burden and worry of planning and providing for retirement. Second, benefits provide a way for employers to create a bond between the firm and the worker. Such a bond and the commitment between the worker and firm that is implied are especially important in firms where workers have (or need to acquire) a significant amount of firm-specific human capital. Employers, who must bear most or all of the cost of investing in firm-specific training, can reap the returns to their investment only if workers remain with the firm over a long period of time. The importance of these two effects has not been quantified convincingly, although there is some evidence that the latter is important (see, for example, the review by Hutchens 1989, or the evidence presented by Topel 1991). But existing evidence suggests that the loss of favorable tax treatment of employee benefits would make it more costly for employers to provide benefits and could indeed lead to social costs in the form of broken job matches that efficiency considerations would suggest should have continued. Section I below briefly discusses the essential features and implications for employee benefits of some of the tax reform proposals that were introduced during the 104th Congress and promise to be considered further in the future. Section II then presents some estimates of how these comprehensive tax reforms would affect (a) the coverage of workers by employer-provided pension and health-insurance plans, (b) employer contributions to pension and health insurance plans, and (c) the shares of compensation received as pensions and health insurance.
